A computational framework for modeling time-frequency distributions

This work presents a computational modeling framework (CMF) for the algorithm formulation of timefrequency tools used in the signal characterization of wireless channels. The work centers on two important original contributions: the development of the CMF itself, which is based on group harmonic analysis and signal algebra operator theory, and the use of Kronecker products algebra (KPA) as an implementation tool-aid in the algorithm mapping of the timefrequency distributions to multicore computational structures. Performance evaluation results are presented which describe the scalability and extensibility attributes the KPA-based algorithm formulations. Implementation performance speed ups are also presented with respect to a PC-based cluster architecture used as a preliminary computational structure. Finally, the work demonstrates how time-frequency methods are instrumental in mapping one-dimensional impulse response functions of wireless channel links in a MIMO system into two-dimensional energy distributions through the use of time-frequency tools.